A. The Feist Decision Did Not Disrupt the Prevailing Intellectual Property Standard Governing the Database Industry.

The proponents of H.R. 2652 suggest that before 1991, the database industry flourished under the sweat of the brow or industrious collection doctrine. This judicial interpretation of the Copyright Act enabled a database publisher to prevent others from copying the content of its database by virtue of the effort it expended in collecting that content. The proponents of H.R. 2652 imply that the Supreme Court destroyed this halcyon environment in 1991 with its Feist decision.

In fact, sweat of the brow was never the prevailing approach for the protection of databases. From the early part of the century, the doctrine existed in tandem with the more mainstream approach of protecting only the expression in the compilation. Sweat of the brow started to decline after Congress overhauled the Copyright Act in 1976. In the 1976 Act, Congress provided that the copyright in a compilation applied only to original selection, coordination, and arrangement of the compiled material. Courts soon recognized that the sweat of the brow doctrine extended beyond the bounds of the new statute. By the early 1980s, the sweat of the brow doctrine was in full retreat. The Feist decision in 1991 was just the final blow. Of course, the Supreme Court went further than just interpreting the Copyright Act, and based its ruling on the copyright clause of the U.S. Constitution.

The fundamental point, however, is that the database industry has been operating without the full benefit of the sweat of the brow doctrine for over twenty years. Ever since the 1976 Act, database publishers have been deciding to create new databases knowing full well that they could not rely on the sweat of the brow doctrine. Moreover, even before 1976, sweat of the brow was not universally accepted. Thus, the proponents of H.R. 2652 are not trying to ameliorate the recent disruption caused by the Supreme Court; rather, they are trying to overturn the status quo of at least the past twenty years.

B. The Feist Decision Has Not Left Databases Vulnerable to Piracy.

Many forms of protection are available to database publishers, most notably copyright. Copyright still applies to the original selection, coordination, and arrangement of the collected material. H.R. 2652's proponents respond that comprehensive electronically stored databases can not meet this standard. Because the databases are digital, they claim, the data are not "arranged" in a traditional sense; instead, the data are "arranged" by the user employing a search engine. Moreover, the compilers make no selection because the databases are comprehensive.

Significantly, the only reported decision involving a comprehensive electronically stored database of something other than phone listings reaches the opposite conclusion. In Corsearch v. Thomson & Thomson, decided in the Southern District of New York the year after Feist, the court considered a database of trademark information Thomson & Thomson had assembled from all 50 states.^{(2) Thomson & Thomson developed a set of fields, and then input the information for each trademark by field. Thomson & Thomson had to correct and standardize the information it received. It also added its own information to each file, such as a code indicating whether the trademark consisted of a word, a design, or a word with a design. The court found that Thomson & Thomson "offered sufficient evidence of its selection, coordination, enhancement and programming of the state trademark data, as well as other contributions that establish the originality and requisite creativity, and thus copyrightability, of the … database."(3)}

The point this case makes is that information in electronic databases typically is not floating around independently waiting to be identified by a search engine. Rather, the author arranges the data in files consisting of linked fields. For a telephone directory, these fields are trivial -- name, address, and phone number. But most commercially valuable databases contain far more fields. The selection of those fields, and the arrangement of bits of data within them, represent at least a minimal level of creativity. To be sure, many of these fields are functionally dictated, yet some reflect the compiler's choice and judgment. Copyright prevents the wholesale copying of such a database. Indeed, copyright prohibits the copying of even a few complete files with linked fields of data, to the extent that the selection of the fields reflects creativity. Therefore, copyright, even after Feist, gives database publishers significant protection.

The proponents of H.R. 2652 counter this argument by citing the Eleventh Circuit's recent decision in Warren Publishing.^{(4) Based on my understanding of the facts, the Eleventh Circuit reached the wrong conclusion in that case. The selection of the primary city in a service area, and the selection of which information to include for each service area, in my opinion constitute protectable expression. Unfortunately, courts make mistakes, particularly in areas as complicated and subjective as copyright. But Congress cannot, and should not, pass a new law every time a lower court makes a mistake. If we were to review every compilation case decided since Feist, I believe that both the proponents and opponents of H.R. 2652 would agree with the result, as a matter of law and equity, in the vast majority of cases. Further, for every case where a court may have extended too little protection, there probably is a case where a court extended too much protection. This is typical of intellectual property law. But looking at all the cases together, copyright provides adequate protection for databases.}

In addition to copyright, database publishers can rely on numerous other forms of protection, including trademark, trade secret, contract, state common law misappropriation, and technological protection. I will not discuss each in detail, but note some recent developments which have enhanced these other forms of protection:

The 104^th Congress passed the Trademark Antidilution Act, which allows holders of famous marks to prevent others from using them even in different product lines where there is no likelihood of confusion. This Act will enhance the value of the names of the established leaders in the database industry.

Judge Easterbrook in the Seventh Circuit recently found a shrink-wrap license prohibiting the copying of telephone listings stored in a CD-ROM enforceable as a matter of state contract law and not preempted by the Copyright Act.^{(5) Although this decision has generated controversy, it is the law within the Seventh Circuit.}

Judge Winters in the Second Circuit recently reaffirmed the state common law misappropriation cause of action. ⁽⁶⁾

The World Intellectual Property Organization Copyright Treaty requires signatories to take adequate measures to prevent the circumvention of copy protection technologies for the purpose of infringement. This Subcommittee, of course, is considering how to implement this obligation. Either approach on the table -- H.R. 2281 or H.R. 3048 -- will significantly enhance database publishers' ability to prevent infringement.⁽⁷⁾

Just like the copyright law, none of these forms is perfect, and not every form of protection will apply in every circumstance. Still, at least one form will apply in virtually all circumstances, and often more than one will apply.

Finally, for an online database whose value derives from its constant updating, the publisher has an easy remedy if it discovers that a subscriber is making unauthorized copies. It can simply discontinue service to that subscriber, and thereby solve the problem. Often self-help works more quickly and effectively than legal relief.

C. There is No Market Failure in the Database Industry Requiring Government Intervention.

Reed-Elsevier and Thomson, the leading proponents of H.R. 2652, retained Laura D'Andrea Tyson, former Director of the National Economic Council, to prepare an economic justification for additional statutory protection for databases. Despite its length, Dr. Tyson's paper lacked what one would expect most from an economic analysis: a serious discussion of the numbers. To be sure, Dr. Tyson included some statistics in an appendix, but the paper itself did not carefully analyze these numbers. And for good reason -- they do not support her conclusion that the database industry needs additional protection.

Dr. Tyson's statistics derive from "The State of Databases Today: 1997," an article contained in the Gale Directory of Databases. The article, prepared by Martha Williams of the University of Illinois, bases its analysis on the Gale Directory itself, which is recognized as the most comprehensive database directory. The most current version of the Williams article, "The State of Databases Today: 1998," shows the following:

Between 1991 and 1997, the number of databases increased from 7637 to 10338, an increase of 35%.⁽⁸⁾

Between 1991 and 1997, the number of files contained within databases increased from 4 billion to 11.2 billion, an increase of 180%.⁽⁹⁾

Between 1991 and 1996, the number of online searches increased from 44.4 million to 79.9 million, an increase of 80%.⁽¹⁰⁾

In short, the database industry has grown remarkably since the issuance of the Feist decision. Not only has the industry created thousands of new databases, but each database on average contains twice as much information. The industry, therefore, is collecting dramatically more information than ever before. Dr. Tyson's response to this evidence of phenomenal growth is that the industry would have grown faster had additional protection been available.

Just as significant as the growth of the industry is its changing structure. In 1977, government, academic, and non-profit entities produced 78% of all databases, while the private sector produced only 22%. By 1991, the government, academic and non-profit share dropped to 30%, while the private sector share increased to 70%. Since 1991, this "privatization" has continued; by 1997, the government, academic, and non-profit share had fallen to 22%, and the private sector share soared to 78%.^{(11) The shift in market share, combined with the increase in the absolute number of databases, indicates that private investment in database creation has increased since Feist. These extremely positive trends fly in the face of the proponents' claims that they need additional government assistance in the form of H.R. 2652.}

Notwithstanding these positive trends, a serious problem does exist in the database industry which does deserve Congressional scrutiny: concentration of ownership. While the total number of databases has increased 35% since 1991, the number of database producers has increased less than 10%.^{(12) Moreover, a handful of the larger database publishers have gone on a buying spree this decade, acquiring other publishers. For example, Reed-Elsevier, which was formed by a merger of a Dutch company and a British company in 1993, purchased LEXIS-NEXIS in 1994, and recently announced its intention to acquire Engineering Information, Inc., Information Handling Services Group, Beilstein Information System, and Wolters Kluwer (itself formed by a merger). Not surprisingly, this increased concentration has led to a significant increase in price. According to figures collected by the Association of Research Libraries, the price of one online service provided by Reed-Elsevier will increase 250% this year, while the price of another Reed-Elsevier compilation will increase 235%.}

Fortunately, it is my understanding that the Antitrust Division of the Department of Justice has initiated an investigation of these Reed-Elsevier acquisitions, as has Directorate-General IV of the European Commission. The House Judiciary Committee should hold hearings on the impact of this concentration of ownership in the hands of a few foreign publishers on competition and creativity in the U.S. database industry. In any event, because of this increasing concentration, the last thing this Subcommittee should do is erect new barriers to entry in the form of additional database protection.

D. The European Database Directive Does Not Necessitate Adoption of H.R. 2652

Proponents of H.R. 2652 argue that Congress must enact additional database protection in order to satisfy the reciprocity requirements of the European Union's Database Directive. In the absence of such a provision, the proponents claim, databases produced by U.S. companies will be vulnerable to predation by European companies. This is an odd argument for Reed-Elsevier and Thomson to make, given their foreign ownership. The argument has no merit, however, regardless of who puts it forth.

While the Directive will deny sui generis protection for U.S. databases if similar protection does not exist here, the Directive will still afford copyright protection to those U.S. databases. And since several countries within Europe currently do not offer adequate copyright protection for databases, the Directive will actually improve matters for U.S. databases relative to the status quo. In other words, U.S. companies will be no worse off than they are now, and indeed might be better off. Given that the proponents have produced no evidence of rampant piracy in Europe at present, they have no grounds for complaining that the Directive will harm them. Of course, the Directive will prevent U.S. companies from extracting substantial parts of European databases. But since the proponents of H.R. 2652 presumably do not engage in such "free-riding" now, this prohibition should not interfere with their activities.

Further, it appears that U.S. companies can receive sui generis protection if they simply establish subsidiaries in Europe. Without doubt, most of the proponents of H.R. 2652 already have European subsidiaries, and thus will automatically receive sui generis protection when the European Union's member states implement the Directive.

II. H.R. 2652 Is A Misappropriation Bill in Name Only

The proponents of H.R. 2652 contend that it addresses the concerns of the opponents of H.R. 3531 introduced by Chairman Moorhead in the 104^th Congress because it is a misappropriation bill as opposed to a sui generis intellectual property bill. From a strictly formal point of view, H.R. 2652 may not create a property right as such. As a practical matter, however, H.R. 2652 grants database publishers as much protection -- indeed, in some respects more protection -- than H.R. 3531.

Moreover, by using the label "misappropriation," proponents of H.R. 2652 create the misimpression that the bill descends directly from the long standing state common law misappropriation doctrine. The following chart compares the most recent articulation of the traditional misappropriation doctrine, NBA v. Motorola, 105 F.3d 841 (2d Cir. 1997), with H.R. 2652. This comparison reveals that H.R. 2652 is far broader than the traditional misappropriation doctrine.

The chart shows three significant differences. First, traditional misappropriation is limited to time sensitive information -- hot news -- while H.R. 2652 applies to collections of information regardless of their age. This means that it applies retroactively, to information already collected without the incentive provided by this bill. It also applies indefinitely into the future. In other words, there is no term limit; a publisher can protect its collection for ever.

Second, the traditional misappropriation doctrine applies only where the defendant competes directly with the collector. H.R. 2652, by contrast, prohibits extraction or use by anyone.

Third, under the traditional misappropriation doctrine, liability attaches only if the defendant's use completely undermines the plaintiff's incentive to create the work. H.R. 2652, conversely, finds misappropriation if there is any harm to an actual or potential market for product. Harm presumably could result from the loss of even one sale.

The bill contains several exceptions and limitations, but they provide little comfort. The exception for non-profit scientific or educational use applies only if the use "does not harm the actual or potential market for the product." This proviso renders the exception useless, given that even one lost sale harms the market for the product.

Because H.R. 2652 applies so broadly, it covers most copyrightable works, but without term limits or copyright's other exceptions. Thus, H.R. 2652 threatens to swallow copyright whole.

III. H.R. 2652 Will Harm OBA Members

H.R. 2652 will harm OBA members both in their capacity as financial institutions and as vendors of Internet products and services. As financial institutions, OBA members produce and use databases. On the producer side, banks develop customer lists which they sell to third parties. They also assemble financial information, such as information concerning the economy or the performance of certain financial products, which they make available to their customers through newsletters or web sites. On the consumer side, banks rely on databases prepared by others for the information they need to make mission critical decisions. For example, every day a bank will study the information contained in a wide variety of databases before deciding what interest rate to charge for various loan products such as mortgages.

Database legislation like H.R. 2652 may enable banks to charge more for the databases they sell, and to prevent other institutions from copying the databases they distribute to customers free of charge. At the same time, H.R. 2652 would increase the cost of the information upon which banks rely, because they would have to pay for information they now receive free or at a nominal cost. Banks have no way of knowing whether the added cost resulting from the legislation would outweigh the benefit, but we suspect it would; banks typically consume more databases than they produce.

H.R. 2652 also would harm OBA members as vendors of Internet products and services. The Internet will achieve its fullest potential as an avenue for electronic commerce if it remains open to all participants and free from domination by one vendor. Recent court decisions have held that copyright does not protect interface specifications, the rules for communication between parts of computer programs. H.R. 2652, however, could grant a developer control over a critical Internet interface specification. The developer could then prevent a product developed by another firm from using these specifications, thereby precluding the product from the Internet market. Accordingly, granting protection for interface specifications could threaten competition on the Internet.

H.R. 2652 does contain an exception for computer programs, but this exception as currently drafted might not exclude interface specifications from the scope of the legislation. At the very least, the Subcommittee should clarify this exception.

The OBA very much appreciates this opportunity to testify, and welcomes any questions the Subcommittee may have.